Mechanical Turk, Amazon's popular crowdsourcing platform, is used by many people and organizations who need repetitive tasks completed. For rapid completion of unskilled (easy) tasks online, crowdsourcing can be an easy and affordable solution. But what if you need people to complete a skilled task? Researchers at ICSI decided to find out.
Here are a few highlights from November 2012 at ICSI:
Our new Swordfish project aims to build speech recognition systems for different languages under severe time and data constraints. Read more about it in our press release.
On the blog, we started a series of posts about common online privacy misunderstandings. So far, there are three posts: part one, part two, part three. This series will continue in December.
We profiled Nelson Morgan (in two parts: read part one, read part two) on the blog. If you are interested in speech research over the past few decades, you'll want to read these.
This is the third in a series of blog posts debunking some common misunderstandings about online privacy. The images used in these posts have been adapted from a tutorial given by Gerald Friedland at ACM Multimedia in Nara, Japan in October.
Most sites are viewable by the public, not just the people you are intending to communicate with. Even non-public sites may be sharing some information with, for example, their advertisers. Never assume that information you provide on a Web page is private.
Last year, a question-answering system built at IBM, named Watson, won the game show Jeopardy! The system combines a large database of knowledge with natural language processing abilities. IBM researchers are now investigating how to use Watson in specialized domains, including health care. On November 9, Alfio Gliozzo, a member of the research staff at the IBM T.J. Watson Research Center, gave a talk at ICSI about the future of Watson. We have now made the slides available.
This is the second in a series of blog posts debunking some common misunderstandings about online privacy. The images used in these posts have been adapted from a tutorial given by Gerald Friedland at ACM Multimedia in Nara, Japan in October.
Many people assume that their online communication is private. While one-to-one communication online may be intended as private, if it is not encrypted, it is not private. Whenever information is sent online, there is the possibility that someone other than the intended recipient can view it. Encryption helps prevent the information from being seen by a third party, so sensitive information (such as emails, chat, or video calls containing information that you want to keep between you and the recipient) should always be encrypted.
Kalle Palomäki recently joined ICSI's Speech Group. He is here on ICSI's Finnish visiting program, which is funded by the Finnish Funding Agency for Technology and Innovation through Aalto University and the Helsinki Institute for Information Technology.
This is the first in a series of blog posts debunking some common misunderstandings about online privacy. The images used in these posts have been adapted from a tutorial given by Gerald Friedland at ACM Multimedia in Nara, Japan in October.
On October 25, Gaël Richard, a professor at Télécom ParisTech, gave a talk on greedy algorithms for representing audio signals. We've now made the slides available. Below is the abstract.
In 1988, ICSI ramped up to full staff, and Morgan began in his role as leader of the Realization Group. The group would focus both on building massively systems and on applications in speech recognition.
The group’s early successes were in designing and building machines powerful enough to do speech recognition. In 1989, the group designed an array of digital signal processing chips in a ring topology that used programmable gate arrays to interconnect processors.